I've been recently asked about how well would Plone fit for a quality system "manual", specially considering the requirements for creating of work instructions.
Some of the requirements would be better structuring of the documents (headings and sections - perhaps with specific workflows), automatic creating of outlines and table of contents for documents, and a nice history view. The paradigm to compare against would be the recent versions of Mediawiki, which has a nice visual editor in-place.
IIRC, Plone is somehow limited on that scope. We had Plone Help Center, to fulfill some of these requirements (I dunno if it is still being updated targeting Plone 5). And standard Plone has some of these features, like the "All content" view (you can create pages in a folder for subsections), and table of contents in pages to autocreate indexes.
But comparing to the equivalent features in Mediawiki, from a user perspective, it seems like Plone is on the losing side.
What are your considerations? Can Plone be sold on that context, or should I recommend adoption of Mediawiki (or other tool)?
Plone is not a Wiki and a Wiki is not a CMS...both systems serve different usecases and both systems are often misused in many bad ways.
Regarding your question: there is no direct answer since your requirements are not completely clear.
In terms of decent technical document Plone is unlikely the right choice when you boil it down to pure content management. Decent technical documentation and re-using content in particular can not be achieved with Plone. The de-facto standard in the field of technical documentation is DITA (an XML dialect widely used in many industries with "topics" as smallest content entity - designed for content-reuse in the first place).
I am working in the field of professional publishing solutions where Plone is a key component as integration plattform but not as CMS in the first place. My solution "XML Director" is capable of dealing with such aspects as you described it but nothing comes for free....systems based on DITA or XML based solutions in particular require careful planning and an analysis of your particular situation. My former "Produce & Publish Authoring Environment" tried to solve your usecase but failed in the end for a variety of reasons. That's why I have XML Director now (as an open and modular platform for publishing solutions in the widest sense).
The cnx.com folks used Plone to create modular textbooks, but that was on Plone 3.x.
It depends on how comfortable you are with Plone too, and if there are other good reasons why you'd use Plone... in which case trying to make Plone work (depending on the specifics of your requirements) could work better. And as Andreas says, a wiki is not going to give you the same level of control (ie. none, really) compared to Plone's security and workflow models.
At UW Oshkosh we put the undergrad and graduate bulletins online using Plone. The "all content" view was customized a bit but that was basically all we needed to do (and train people how to use Plone). Workflow, sharing, and Working Copy Support let departments edit & revise their sections while maintaining control over final edit approvals.
We're working with a Fortune 100 client who is using Plone to build online manuals for several of their hardware and software products. We installed PloneHelpCenter for them, with a few customizations and they are making use of the ReferneceManual content-type. We also developed a BrowserView which can be called on a folderish structure (ReferneceManual or regular Plone Folder) which creates a PDF document of the folder contents, complete with index, cover page, etc. The client is happy with the solution and they have moved several more business units into using it for creating technical manuals.
I never used Sphinx to edit content. If my understanding about it is correct, the content is edited as reStructuredText, right?
If so, and if the content creator need to edit in reStructuredText, that might be an obstacle, because the content creators are non-technical people (at least non-IT-techs) and would be averse to use a non-visual tool to edit. As a matter of fact, Mediawiki is being considered now because it has a nice visual editor.
Understood. So Sphinx adds the burden of requiring knowledge of rst and for non-technical users that's a blocker.
Plone is up for the task. My dream solution would be a system that "bakes" the finished manuals into Sphinx for deployment.